Why a Confocal Sensor Cannot Measure on a Highly Tilted Surface

The mechanism is aperture walk-off

A chromatic confocal sensor focuses a broad spectrum along the optical axis and determines distance from which wavelength returns through a small confocal aperture. When the surface is tilted, the reflected beam leaves at twice the tilt angle relative to the axis. By the time it travels back to the aperture, the spot has moved sideways by an amount that grows with both tilt and working distance. Once the spot no longer overlaps the aperture, the returned signal drops and the measurement becomes noisy or fails. This is a geometric effect and it affects every confocal sensor, regardless of make.

Why the limit depends on working distance, not only on tilt

The lateral displacement is proportional to the distance the beam travels, so a sensor operating further from the part tolerates less tilt. This is counter-intuitive: users often assume a longer range sensor is more forgiving, but its larger standoff consumes the angular budget faster. When a measurement on a tilted surface is marginal, moving to a shorter standoff variant usually helps more than anything else.

What the acceptance angle means in practice

Manufacturers quote a maximum tilt, typically expressed as a half-angle or as a slope. That figure assumes a perfectly reflecting surface and a perfectly aligned sensor. Real parts have roughness, oxidation or a protective film, which reduces returned energy before the geometry test is reached. On real parts the usable tilt is therefore usually well inside the quoted figure, and the practical way to find it is to measure a step artefact at increasing tilt until the reading becomes unstable.

The four ways to recover a tilted measurement

First, reduce the standoff by selecting a shorter-range variant if the measurement range allows. Second, rotate the sensor so the tilt is in the plane in which the sensor is most tolerant, which matters because the aperture geometry is not always rotationally symmetric in the same way. Third, split the measurement: measure the tilted face as a series of profiles and fit the surface, rather than trying to measure it point by point. Fourth, change the fixture so the surface is normal to the sensor at the measurement point, which is often cheaper than changing the sensor.

Why tilting the sensor is usually better than tilting the part

Tilting the part changes the relationship between the part and every other inspection station, and it changes the datum. Tilting the sensor keeps the part datum intact and changes only the optical path. In a line with several stations, sensor-side tilt is therefore the option that does not invalidate the rest of the process.

What to specify when the part cannot be re-fixtured

State the maximum surface slope at the measurement point, the required range, and the required standoff together. These three determine whether any confocal sensor can work at all: if the slope is large and the standoff is forced to be long by the fixture, the honest answer may be that a different measurement principle is required, and that should be established before a sensor is purchased rather than after.

Frequently asked questions

Why does a confocal sensor lose signal when the surface is tilted?

The reflected beam leaves at twice the tilt angle and walks off the small confocal aperture, so returned energy drops.

Does a longer working distance tolerate more tilt?

No, less. Lateral displacement grows with the distance the beam travels, so a longer standoff consumes the angular budget faster.

Can I exceed the quoted maximum tilt on a real part?

Usually not. The quoted figure assumes a perfect reflector; roughness and films reduce returned energy before the geometric limit is reached.

What is the cheapest way to recover a tilted measurement?

Re-fixture so the surface is normal at the measurement point, or rotate the sensor rather than the part so the datum is preserved.

Should I tilt the part or the sensor?

The sensor, whenever possible. Tilting the part changes the datum and affects every other station in the line.

When should I abandon confocal for a tilted surface?

When the required slope and the fixture-forced standoff cannot both be satisfied. Establish this before purchase, not after.